Waveform Inversion in Ultrasound Tomography from Homogeneous Starting Models

While waveform inversion is the clinical gold standard for obtaining high-resolution images of sound speed and attenuation in tissue, it was long assumed that bent ray tomography was necessary to obtain a starting model, especially with a ring array transducer. However, recent works show that waveform inversion is possible with a reasonable choice of homogeneous starting models and sufficiently low-frequency content in the waveform to avoid cycle skipping. In the case of transcranial imaging, we cannot afford to use a bent ray tomography because it quickly becomes unstable in the presence of a highly refractive skull. We examine the parameters that would enable waveform inversion to image the soft tissue in the brain through the high sound speed and attenuation in the skull.